GitLab

newsize is a long, but storage is an int. This means the allocation
could succeed but storage would overflow.

Closes #2300

The link pointed to the wrong structure documentation.
Looks like a cut-and-paste error.

Signed-off-by: Ralph Giles <giles@thaumas.net>

Gitlab now supports a `default` section on the build description,
which sets defaults which can be overridden by specific jobs.

This encapsulates things a little better and avoid duplicating
runner tags.

Specify an explicit version of the base container image for more
repeatable results. Note that this will need to be bumped periodically.

We've moved hosting to a gitlab instance. Make use of the built-in
integration testing framework to verify commits.

This has parallel jobs to run the autotools and cmake builds.

Note that the autotools job needs cmake because the `distcheck`
target verifies that the generated source package also builds
correctly with cmake, in particular to check that the necessary
files are included.

This description of the build is old and unmaintained. It's not
useful to have untested code in the tree.
Signed-off-by: Thomas Daede <daede003@umn.edu>

This description of the build is very old, unmaintained, and
not integration tested. I don't think it's useful to users,
who can use the supported autotools or cmake builds, or just
add the few source files directly to their own projects.
Signed-off-by: Thomas Daede <daede003@umn.edu>

Close the C4244 warning while compiling under windows with cmake+msbuild.
Signed-off-by: Tristan Matthews <tmatth@videolan.org>

This effectively reverts f8ce071e which was
probably broken by 6449883c.

add `make test` to README.md

On windows platform, building with MinGW can run 'make check' (autoconf)
and 'ctest' or 'make test' (cmake), while building with MSBuild can only
run 'ctest -c Release|Debug'.
Signed-off-by: Ralph Giles <giles@thaumas.net>

These should match http://downloads.xiph.org/releases/ogg/SHA256SUMS

Result of running make -C doc/libogg update-doc-version.

List of significant changes since the last release tag.

Describe how to update documentation versioning. No need for
sed one-liners each time.

Include everything in the cmake directory in the `make dist`
target. Fixes a missing include error in the CMake build
from a release tarball for `CheckSizes.cmake`.
Signed-off-by: Tristan Matthews <tmatth@videolan.org>

Verify that the CMake build works from a release tarball.
This will give coverage for problems with part of the
CMake build description not being included in the
source package. We still consider the `make dist`
output from the GNU autotools build to be canonical.

NB CMakelists was in the repo at the time of the 1.3.3 release,
but didn't make it into the 1.3.3 dist target, so it can't be
said to be available for that release.

cmake in build dir is recommand by officially.
Signed-off-by: Ralph Giles <giles@thaumas.net>

Use the correct library filename and only package the dll
if the build produced one.
Signed-off-by: Ralph Giles <giles@thaumas.net>

The 2015 version of the project files can be opened by newer
versions of Visual Studio, and we want to deprecate the project
files in favour of the CMake build, so it doesn't make sense
to support three different versions.
Signed-off-by: Ralph Giles <giles@thaumas.net>

Use <sys/types.h> not <inttypes.h> when building for Apple platforms,
as it works with modular dynamic-framework builds.
Signed-off-by: Ralph Giles <giles@thaumas.net>

Copy the struct definition from ogg.h which has better comment
alignment.

The pageno is declared `long` in the header, not `int`, although
it's a 32-bit value in the page header.

Thanks to Spencer Russell for reporting the issue.

oy.data memory is alloc by ogg_sync_buffer(), but does not call free()
before main() exit. After fixing it, I test test_framing by valgrind.
And no more memory leak in test_framing and test_bitwise right now.
Signed-off-by: Ralph Giles <giles@thaumas.net>

According to VS2017 docs, the initial LIBRARY line must include
a dll name. MSVC handles the missing name, but MinGW does not.

Removing the line entirely is also accepted by both toolchains,
but is contrary to documentation.

Based on a patch by evpobr in https://github.com/xiph/ogg/pull/42
No ABI change according to comments there.